28 September 2013

Locating and Repairing Bad Pixels

If you have pixels - be they nodes, modules or strip - at some point you are going to have a failure.  That failure could be when you purchase them or over a period of time due to physical damage or other causes.  So, you'll need to have the ability to locate and then replace the bad pixel.  This video shows how to perform this process on nodes (8mm / 12mm) but it also is the exact same process if you have any other type of pixel. 


As mentioned in the video, if you do pixels, you'll need to solder and it's not nearly as complex as you might expect.  Check YouTube for videos on how to solder.

5 comments:

  1. could a bad pixel all show as one that flashes to white or another color, instead of the steady 3 color change test pattern, and all the ones after it also flash randomly.

    For example - 134 node setup - 3 power supplies
    try all 134 - w tiny pix and 3 power supplies - most flash to white randomly while trying to change colors

    if I skip the first 34 (just because that is the next 2811 connector) - the other nodes work fine. Color change in test sequence. And even work with a DMX test.

    So narrowing it down. I try just the first 34 nodes alone (and disconnect the rest)
    the first 4 pixels cycle colors in the test mode just fine.
    the 5th one flash to white every once in a while
    and all the rest of the 34 flash to white all the time randomly (while trying to change colors in the test sequence)

    is it likely a bad 4th or 5th pixel?
    or some other issue?

    -Bruce

    ReplyDelete
    Replies
    1. If you have a string of pixels and it doesn't seem there is a clear point where the problem is occuring (such as all the pixels following the working ones), I'd recommend cutting the string in half and then testing both halfs. Then keep performing that process until you locate the bad pixels (cut in half, cut the remaining bad section in half, etc.)

      Delete
    2. I cut after node 4.
      Then attached an extra 2811 3pin end temporarily - and hooked up my extra 16 nodes to these 4 nodes
      hooked up tiny pix w one power supply
      worked fine - demo rgb blue

      so I cut out node 5
      Soldered the line back together (in the field on the display, w srwap on each wire, and then a large swrap over all 3)

      Worked in demo test
      now working via DMX test sequence
      I am letting it burn in test for an hour on the DMX sequence

      I have a video of the flashing to full white randomly symptom to share.
      (to help others in case they see this symptom on their fire pix string)

      -Bruce

      Delete
  2. So to wrap up my lessons learned. After cuttting out nodes, 3, and 5.
    A week into this process, Node 9 finally failed.
    And the kicker - when it failed - it clearly shorted the data channel and all the nodes after it were full on (white).
    So I replaced this node (I did not have an extra nodes left in this part of my tree), and ...
    everything has been running fine for a week. (minus one node that does not tranistion from white to blue quickly, and flashes red a bit for quick changes, but that is minor), as it works fine for other sequences - and since things are stable I am going to leave it alone)

    So the Summary:
    Use the tiny pix and first 50 nodes at halloween. Noticed odd programming setting in the tin pix. Junk data setting for demo.

    Then built the tree.
    Would run ok.
    Then not run.
    Then all nodes would random flash to white while trying to demo test.
    So I started cutting out nodes where the flashing started.
    5,3
    Then more tiny pix odd behavior, it would only work with 50 nodes (for me 50 nodes did not end at a connector, so that was not the issue)
    Then tiny pix would not work at all. Checking the programming showed it had set the ghost pixels to 255.
    Then node 9 failed. With all the nodes after it on full white.
    I replaced node 9. (since I need 16 nodes in a leg of my tree)
    demo test RGB worked fine.
    DMX testing has worked fine for a week.
    I have not checked the tin pix programming again (since it is outside and working)

    So if you see symptoms like these
    odd tiny pix programming
    works then does not
    flashes to white
    works, then only works with 50, then none (programming)

    My un-official opinion is that you have a node doing bad things to the data line. (and thus causing issues for the tiny pix and the other nodes)

    -Bruce

    ReplyDelete

We welcome all feedback